AutoFL: A Tool for Automatic Multi-granular Labelling of Software Repositories

2024-08-05

arXiv.org

AI Summary

Software developers face inefficient and time-consuming challenges in comprehending large, multifunctional codebases; existing README-based, coarse-grained project-level categorization fails to support fine-grained functional understanding. To address this, we propose AutoFL—the first automated, cross-granularity functional domain labeling method supporting file-, package-, and project-level annotations without relying on non-code documentation (e.g., READMEs). AutoFL directly models source code semantics via a weakly supervised learning framework that integrates code text parsing, multi-granularity semantic embedding, and hierarchical aggregation for end-to-end label generation. Evaluated across multilingual open-source projects, AutoFL significantly improves the accuracy, consistency, and interpretability of functional labels compared to baselines. It effectively alleviates key bottlenecks in software comprehension by enabling precise, scalable, and documentation-agnostic functional awareness.

Technology Category

• Natural Language Processing: Code Generation / Program Synthesis from Natural Language
• Machine Learning: Transfer, Domain Adaptation, Multi-Task Learning
• Application Domains: Software Engineering

Application Category

• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Search and Retrieval-Augmented AI: Multilingual and cross-lingual Web search
• Economics, Online Markets and Human Computation: Humans versus LLMs for data annotation and labeling

Abstract

Software comprehension, especially of new code bases, is time consuming for developers, especially in large projects with multiple functionalities spanning various domains. One strategy to reduce this effort involves annotating files with meaningful labels that describe the functionalities contained. However, prior research has so far focused on classifying the whole project using README files as a proxy, resulting in little information gained for the developers. Our objective is to streamline the labelling of files with the correct application domains using source code as input. To achieve this, in prior work, we evaluated the ability to annotate files automatically using a weak labelling approach. This paper presents AutoFL, a tool for automatically labelling software repositories from source code. AutoFL allows multi-granular annotations including: extit{file}, extit{package}, and extit{project} -level. We provide an overview of the tool's internals, present an example analysis for which AutoFL can be used, and discuss limitations and future work.

Problem

• Automate labeling software repositories for faster code comprehension
• Enable multi-granular annotations at file, package, and project levels
• Improve domain-specific file classification using source code input

Innovation

• Automated multi-granular labelling of repositories
• Uses source code for domain-specific annotations
• Supports file, package, and project-level granularity